US bars DeepSeek technology from most civilian government systems
Developed over time first seen 19 hours ago
The US Department of Homeland Security has ordered DeepSeek AI technology removed from most civilian government systems, including tools built with its models. The ban adds to Washington’s scrutiny of Chinese AI firms amid competition with China over the technology’s development and use.
The order was made public on Monday but took effect at the end of July. It applies to civilian agencies including Treasury, State and Justice, while excluding the Department of Defense; the notice did not specify the security concerns. DeepSeek’s open-source models are popular partly because they cost less: the company said training its first model cost about $6 million, compared with roughly $100 million for OpenAI’s GPT-4.
- US civilian agencies must remove DeepSeek technology.
- The ban excludes the Department of Defense.
- DeepSeek said its first model cost about $6 million to train.
New here? Start with this
DeepSeek is a Chinese artificial intelligence company whose models can generate text and help with other tasks. Its tools are open source, meaning developers can use and adapt them, and have attracted attention partly because they are relatively inexpensive.
The US federal government is divided into civilian agencies, such as the departments of Treasury, State and Justice, and the Department of Defense, which is responsible for the armed forces. The order applies to civilian agencies and also covers technology built using DeepSeek models.
The restriction comes amid wider US scrutiny of Chinese technology firms and competition between the two countries in artificial intelligence. US officials and companies have accused some Chinese firms of copying American AI models; the public notice did not set out specific security concerns behind this order.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
A government-wide restriction can be justified as prudent protection of sensitive systems when officials cannot be confident that foreign-developed AI tools meet security and data-handling requirements. Limiting use across civilian agencies may reduce exposure while those risks are assessed, especially given the strategic competition over AI between the US and China.
The case against
A broad ban may be disproportionate if the order does not identify specific security risks or distinguish among models and uses. DeepSeek’s open-source tools are relatively inexpensive, so excluding them could raise costs, narrow agencies’ options and hinder access to useful technology; targeted safeguards and case-by-case review could address risks while preserving those benefits.
More coverage
AI Americas Art Asia Culture Government Politics Technology World
Read the full article at the source →
Originally published by BBC Technology as “US bans DeepSeek AI from government use as tech tension with China continues”.